A cohort of patients with chronic type B hepatitis is being evaluated and followed to determine the long-term natural history of this common form of chronic liver disease. Selected patients have been entered into the therapeutic trials in which antiviral or immunomodulatory agents have been administered. A randomized controlled trial of a four month course of alpha interferon in 45 patients with chronic type B hepatitis has recently been completed. In this trial, patients were randomized to receive either (A) 5 million units sc of interferon sc daily for four months, (B) 10 million units sc of interferon every other day for four months, or (C) no therapy. All patients have been followed for one year and control patients have been "crossed over" to receive a four month course of interferon after evaluation following the initial year of follow-up. During the period of therapy, 10 of the 31 treated (32%) but only one of 14 (7%) untreated patients lost serological markers of active hepatitis B viral replication (serum HBV-DNA and DNA polymerase activity). This difference was not statistically significant (p = .09). Among 10 control subjects crossed over to treatment, 4 (40%) responded with clearance of HBV-DNA from serum. Thus, the response rate to alpha interferon alone, given for a four month period, is low. Analysis of factors that might predict whether patients would respond to alpha interferon in the controlled trial as well as in our other studies of interferon therapy for this disease demonstrated that female sex and height of serum aspartate aminotransferase activity (AST:SGOT) were the two best predictors of a favorable response.